首页> 外国专利> DEVICE, METHOD AND PROGRAM OF STOCHASTIC CONTEXT-FREE GRAMMAR INQUIRY

DEVICE, METHOD AND PROGRAM OF STOCHASTIC CONTEXT-FREE GRAMMAR INQUIRY

机译:随机无上下文语法查询的装置,方法和程序

摘要

PPROBLEM TO BE SOLVED: To provide a technique for reducing a calculation cost while maintaining the accuracy of an inquiry result in a stochastic context-free grammar inquiry. PSOLUTION: In the stochastic context-free grammar inquiry, it is assumed that grammar G to be defined by an aggregate of terminal symbols, non-terminal symbols (nodes) and generation rules with probabilities is given. Firstly, a method of a stochastic context-free grammar inquiry rearranges the grammar G at random (Step S101). Next, the method puts together all the nodes to calculate an approximate probability P2 (step S105), puts together the nodes for each height to calculate an approximate probability P1 while pruning (pruned parts are excluded from the objects of the inquiry) the nodes for the grammar G (Step S107), compares the result with a probability of a solution candidate (Steps S106, S108), and performs accurate probability calculation after a reduction in the number of grammars and nodes to be used for calculation of an accurate probability P while pruning the nodes (Step S109). PCOPYRIGHT: (C)2010,JPO&INPIT
机译:

要解决的问题:提供一种在保持随机上下文无关文法查询的准确性的同时降低计算成本的技术。

解决方案:在无上下文随机语法查询中,假定语法G由终端符号,非终端符号(节点)和具有概率的生成规则的集合定义。首先,随机无上下文语法查询的方法随机地重新排列语法G(步骤S101)。接下来,该方法将所有节点放在一起以计算近似概率P2(步骤S105),将每个高度的节点放在一起以计算近似概率P1,同时对节点进行修剪(修剪掉的部分从查询对象中排除)。语法G(步骤S107),将结果与候选解的概率进行比较(步骤S106,S108),并且在减少用于计算准确概率P的语法和节点的数量之后,进行准确概率计算。在修剪节点的同时(步骤S109)。

版权:(C)2010,日本特许厅&INPIT

著录项

  • 公开/公告号JP2010066970A

    专利类型

  • 公开/公告日2010-03-25

    原文格式PDF

  • 申请/专利权人 NIPPON TELEGR & TELEPH CORP NTT;

    申请/专利号JP20080232180

  • 发明设计人 FUJIWARA YASUHIRO;

    申请日2008-09-10

  • 分类号G06F17/27;

  • 国家 JP

  • 入库时间 2022-08-21 19:03:44

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号